

Zero-Knowledge AI: The Frontier of Cryptography | Zuzalu #2
Jul 5, 2023
Join AI researcher Daniel Shorr from Modulus Labs and cryptographer Justin Drake from the Ethereum Foundation as they dive into the exciting intersection of zero-knowledge technology and artificial intelligence. They discuss how ZKML can enhance the integrity of AI outputs and the transformative potential of on-chain verification. With insights on decentralized applications and AI in gaming, they explore the risks of biased models and the promising future of reliable and secure AI systems, urging listeners to embrace this new frontier.
AI Snips
Chapters
Transcript
Episode notes
ZK Proof Properties
- Zero-knowledge proofs (ZK) verify computations without redoing them.
- This is useful for compressing information and maintaining blockchain security.
Stress-Testing ZK
- ZK proofs can scale blockchains.
- Applying them to machine learning models stress-tests this capability.
ZKML's Purpose
- Many AI models will likely emerge, generalizable to any purpose.
- ZKML can verify these models on-chain, ensuring authenticity.